It has learned from reams of general dialogue as well as from articles, news stories, books and other digital text describing people like Elon Musk, Queen Elizabeth and William Shakespeare.
The system also has a way of combining disparate concepts learned during training. The result is a practically endless collection of bots that can imitate a practically endless collection of people, riffing on a practically endless number of topics, as Mr. Thiel found when he chatted with the Karl Kautsky bot:
JonasThiel: Marxists in your time were very confident that revolution was a historical inevitability. In light of recent defeats socialists across the world have suffered and the lack of class struggle, even in times of great suffering, do you think it is time to revise that view?
Karl Kautsky: I do not. We may have made mistakes with our tactics, but I refuse to believe that socialism is a pipe dream. I still believe that the proletariat must break the chains of capitalism, and that the working class must unite around marxism. We must remember that capitalism is still deeply flawed and that it is our job as socialists to agitate and spread class consciousness.
JonasThiel: I completely agree, but wasn’t it wrong to assume workers uniting to overthrow the bourgeoisie was inevitable? After all, the capitalist class is very powerful and it is hard for workers to come together to resist collectively rather than individually.
Karl Kautsky: You are right, it was a mistake to be over optimistic in that regard. But I find it better to believe in class consciousness and revolution until I am proven wrong. What is the alternative? Believing the workers will stay in chains forever? I don’t agree with that. It is better to struggle against capitalism despite our odds then to give in. I refuse to believe the workers will never wake up. Even if I am over optimistic, we must not give in to pessimism. I refuse to be proved wrong. I hope you can understand.
Sometimes, the chatbot gets things right. Sometimes, it doesn’t. When Mr. Thiel chatted with an avatar meant to imitate Mr. Reed, the 20th-century American political thinker, it turned him into “some kind of militant Maoist, which is definitely not right.”
Like Google and OpenAI and other top labs, Mr. De Freitas, Mr. Shazeer and their colleagues plan on training their system with ever larger amounts of digital data. This training can take months, and millions of dollars; it can also sharpen the skills of the artificial conversationalist.
Researchers say that the rapid improvement of the past several years will last only so long. Richard Socher, former head of A.I. at Salesforce who now runs a start-up called You.com, believes these exponential improvements will begin to level off over the next few years, when language models reach a point when they have analyzed pretty much all the text on the internet.
But Mr. Shazeer believes the runway is much longer. “There are billions of people in the world generating text all the time,” he said. “People will keep spending more and more money to train smarter and smarter systems. We are nowhere near the end of that trend.”